Article Text
PDF
Abstract
Background Indeterminate readout of the quantitative interferon-γ release test (QFT) for Mycobacterium tuberculosis screening is a specific laboratory finding for systemic lupus erythematosus (SLE), which may be due to T-cell exhaustion and abnormal programmed death receptor 1 (PD-1)/programmed death-ligand 1 (PD-L1) signalling.
Methods We enrolled 104 patients with SLE and 225 with other rheumatic musculoskeletal diseases (RMDs) who presented to the outpatient clinic between 2020 and 2023. Twenty healthy donors served as the controls. The QFT was performed in all participants, and those with indeterminate results were compared among the groups. Immunophenotyping and functional assays were performed using blood mononuclear cells. Interferon (IFN)-γ was detected in vitro and ex vivo in patients with SLE with indeterminate or negative QFT results, before or after rituximab therapy.
Results 104 patients with SLE had a significantly higher rate of indeterminate QFT results was significantly higher (17.31%) than that of 225 patients with RMD (3.56%). Patients with SLE with indeterminate QFT had more active disease (SLEDAI-2K, mean 10.94 vs 4.02, p<0.0001), including a higher incidence of active nephritis (55.56% vs 29.07%). Indeterminate QFT in SLE is mainly caused by an insufficient IFN-γ response in CD8+T cells with exhausted immunophenotypes. The abnormal interaction between exhausted PD-1 high CD8+ T cells and activated PD-L1 low memory B cells in SLE can be reversed with a PD-1 agonist or increased PD-L1 expression. Rituximab treatment indirectly reversed this IFN-γ response.
Conclusion The PD-1/PD-L1 signalling pathway, which governs the crosstalk between exhausted CD8+ T cells and activated memory B cells, is a mechanistic explanation for insufficient interferon-γ response in patients with SLE.
- Lupus Erythematosus, Systemic
- Cytokines
- B-Lymphocytes
- T-Lymphocyte subsets
- Rituximab
Data availability statement
Data are available upon reasonable request. Not applicable.
This is an open access article distributed in accordance with the Creative Commons Attribution Non Commercial (CC BY-NC 4.0) license, which permits others to distribute, remix, adapt, build upon this work non-commercially, and license their derivative works on different terms, provided the original work is properly cited, appropriate credit is given, any changes made indicated, and the use is non-commercial. See: http://creativecommons.org/licenses/by-nc/4.0/.
Statistics from Altmetric.com
WHAT IS ALREADY KNOWN ON THIS TOPIC
The quantitative interferon-γ release test (QFT) for systemic lupus erythematosus (SLE) yielded a significantly higher percentage of indeterminate results.
WHAT THIS STUDY ADDS
Patients with SLE and indeterminate QFT had more active disease and higher incidence of nephritis.
The programmed death receptor 1 (PD-1)/programmed death-ligand 1 signalling pathway between exhausted CD8+ T cells and active memory B cells causes indeterminate QFT in patients with SLE.
HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY
PD-1 agonists represent an appealing approach for modulating CD8+ T-cell exhaustion in active SLE.
Introduction
Systemic lupus erythematosus (SLE) is a complex autoimmune disease that causes severe immune dysfunctions. Extensive immunosuppressive or biological treatment is often required to restore immune system balance in patients with SLE.1 Interferon-γ release assays (IGRA) for Mycobacterium tuberculosis (TB) screening are becoming routine clinical practice conducted prior to the initiation of extensive immunosuppression, especially in TB epidemic regions.2 3 Quantitative IGRA, such as QuantiFERON TB (quantitative interferon-γ release test (QFT)), have been widely used to measure interferon (IFN)-γ production in peripheral blood cells stimulated with TB-specific antigens (Early Secreted Antigenic Target-6 and Culture Filtrate Protein-10).
It is noteworthy that indeterminate QFT4 result are prevalent among patients with SLE in juxtaposition with other rheumatic musculoskeletal diseases (RMDs),5 which is largely attributed to suboptimal readouts of the positive control according to our preliminary observations. In other words, peripheral blood mononuclear cells (PBMCs) from patients with SLE have shown an insufficient interferon-γ response (IIR) to phytohaemagglutinin (PHA), which might lead to this interesting laboratory finding, that is, indeterminate QFT results.
Co-stimulatory and co-inhibitory signals regulate the interaction between T and B cells during the inflammatory response, which is crucial for SLE pathogenesis. The activation of autoreactive T and B cells in SLE impairs the co-stimulation and co-inhibition pathways. As part of pathogenesis, the defective checkpoint molecules expressed on the surface of T and B cells, such as cluster differentiation (CD)40/CD40L and programmed death receptor 1 (PD-1)/programmed death-ligand 1 (PD-L1), aggravate T-cell exhaustion in SLE.6 7 This study aimed to address the immunological mechanisms underlying this relatively SLE-specific in vitro phenomenon. It has been hypothesised that T-cell exhaustion, a senescent T-cell phenotype, is a key mechanistic driver. Thus, in-depth immunophenotyping and functional experiments were performed to investigate IIR in SLE, with an emphasis on exhausted T cells.
Patients and methods
Patients
From January 2020 to January 2023, 329 patients with RMD and 20 healthy donors, excluding those with active TB infections, were enrolled at the Department of Rheumatology, Renji Hospital, Shanghai. A total of 104 patients fulfilled the Systemic Lupus International Collaborating Clinics 2012 classification criteria for SLE.8 Disease activity was calculated according to the SLE Disease Activity Index (SLEDAI-2000)9 at sampling.
Quantitative IFN-γ releasing test for TB
The assay was performed using a commercially available kit, according to the manufacturer’s instructions (Zhuhai Lizhu Reagent, China). To gauge the TB-specific IFN-γ response (ESAT-6 and CFP-10, TB specific antigen, QFT-T tube), the QFT had both a negative and positive control, with ‘nil’ (medium, negative control, QFT-N tube) and mitogen PHA (positive control, QFT-P tube) as stimuli. After 16–24 hours of incubation for the three aliquots of heparinised whole blood at 37°C in a humidified atmosphere, the supernatant was separated from each tube after centrifugation at 3000 relative centrifugal force for 15 min. The amount of IFN-γ was measured using chemiluminescence (LEACL-600, Lizhu, China). The results of QFT were considered positive if the value of T minus N IFN-γ was ≥0.25 IU/mL and ≥N/4. The results were considered negative if the IFN-γ level of T minus N was <0.25 IU/mL and the P minus N value was ≥0.4 IU/mL. The results were defined as indeterminate if the IFN-γ value of T minus N was <0.25 IU/mL and P minus N was <0.4 IU/mL, or if the IFN-γ value of T minus N was ≥0.25 IU/mL but below N/4. If the IFN-γ level in QFT-N is greater than 8.0 IU/mL, the result is also classified as indeterminate regardless of the value of QFT-T or P.
Serum IgG purification and ELISA for anti IFN-γ IgG antibody
IgG from SLE sera was purified using a HiTrap Protein G HP column (GE HealthCare, USA) and quantified using a BCA Protein Assay Kit (Sigma-Aldrich, USA). The neutralising anti IFN-γ antibody in the plasma was detected using ELISA. The 96-well plates were coated with 2.5 µg/mL recombinant IFN-γ protein (Hangzhou Qitai Biotechnology, China) and incubated overnight at 4°C. Discard The coating solution, the plates were blocked with 1% Bovine Serum Albumin solution at 37°C for 2 hours and washed three times with Tris-HCl buffer with tween-20. 100 µL purified IgG in different concentration gradients (1–5 µg) of patients with SLE was added to 96-well plates and incubated at room temperature (RT) for 30 min. The plates were then washed thrice with TBST. Then, 100 µL of horseradish peroxidase-mouse anti-human neutralising IFN-γ IgG (1:50 000, CST, USA) was added and incubated under light for 30 min. After washing with Phosphate buffer added with tween-20 three times, 3,3',5,5'-tetramethylbenzidine buffer was added and the reaction was stopped with 0.5% H2SO4 for 10 min. Absorbance was measured at 495 nm.
Peripheral blood mononuclear cells isolation and lymphocyte sorting
PBMC were separated from buffy coats using a density gradient (Ficoll-Hypaque, TBD, China) by centrifugation at 1200×g for 10 min. PBMC were purified into CD4 and CD8 positive cells using a magnetic cell sorting system (Miltenyi Biotec, Bergisch Gladbach, Germany) according to the manufacturer’s instructions. The purity of the isolated CD4+ and CD8+ T cells was >95%.
Lymphocyte culture
The 24-well plates were added to PBMC (1×105/ well), purified CD4+ and CD8+ T cells (1×105 cells/well) and cultured for 24 hours with 1 µg/mL of anti-CD3/CD28 antibodies (BD Biosciences, USA) in complete Roswell Park Memorial Institute 1640 medium in the absence or presence of different concentrations of PHA. PBMC (1×105 cells/well) and PD-1 agonist (50 µg/mL) or PD-1 inhibitor (30 µg/mL) monoclonal antibodies donated by ZhenRong Biotechnology (Shanghai, LTD) were co-cultured in the presence of anti-CD3/CD28 for 72 hours. Cultured cells were used for total RNA extraction and flow cytometry.
RNA extraction and complementary DNA synthesis
Total RNA from the cultured cells was extracted using TRIzol reagent (Invitrogen, USA) and reverse-transcribed to generate complementary DNA (cDNA) using a cDNA Reverse Transcription kit (TaKaRa, Dalian, China), according to the manufacturer’s instructions. The converted cDNA was quantified by absorbance at 260 nm using a NanoDrop (NanoDrop Technologies, Wilmington, Delaware, USA) and stored at −80°C for subsequent analyses.
PCR amplification of IFN-γ and immune checkpoint genes
The gene list and primer sequences are shown in online supplemental table 1. Quantitative real-time PCR was performed using an ABI 7500 Fast Real-Time PCR System with SYBR Green Mix (TaKaRa, Dalian, China). Reactions were prepared in a total volume of 20 µL, with 2 µL cDNA, 10 µL SYBR green mix, 7 µL of DEPC water and 1 µL of 5 µM forward and reverse primer mix in a 384-well plate (Applied Biosystems, Warrington, UK). The relative gene expression is specified as the corresponding 2ˆ(−ΔCT) value.
Supplemental material
Intracellular IFN-γ staining
The washed PBMC stimulated by Mhorbol Ester/lonomycin (Absin, China) for 6 hours were stained with 20 µL surface antibodies, anti-PerCP-CD3, anti-FITC-CD4 and anti-PE-CD8 (BD Tritest, USA) at 4°C for 20 min. After washing, the samples were fixed and permeabilised (BD Perm/Wash buffer, USA) and 10 µL of Allophycocyanin-conjugated monoclonal mouse anti human IFN-γ (Becton Dickinson, USA) was added for 15 min in the dark. The cells were washed and analysed using a Calibur analyser (BD Biosciences, USA). For each analysed sample, a minimum of 5000 events were recorded. All data were analysed using FlowJo V.7.6.
Flow cytometry staining for T-cell exhaustion markers and B-cell subsets and high-dimensional analysis
Antibodies purchased from BD Biosciences (USA) to surface markers were used to generate an eight-colour flow cytometry panel: PerCP-Cy5.5-CD3, APC-H7-CD4, Alexa Fluor 700-CD8, PE-Cy7-PD-1, Alexa Fluor647-LAG3, BV786-TIM3 PE-CTLA-4, BV711-BTLA and BV480-KLRG1. The staining panel of B-cell subsets was as follows: Transitional B cells (CD19+CD24+CD38+), naïve B cells (CD19+CD24+CD27–IgD+), class-switched-memory B cells (CD19+CD24+CD27+IgD–), non-class-switched-memory B cells (CD19+CD24+CD27+IgD+), double negative B cells (CD19+CD24+CD27–IgD–), q1Plasma cells (CD19dimCD24–CD38highCD27highCD138+) and plasmoblasts (CD19dimCD24–CD38highCD27highCD138–). PBMC were stained for 30 min on ice with specific antibodies in Brilliant Stain Buffer (BD Biosciences) according to the manufacturer’s instructions. Following incubation, the cells were washed thrice with cold phosphate-buffered saline (PBS) and centrifuged at 3000×g for 5 min. Subsequently, the cells were resuspended in PBS. Samples were acquired on a BD FACSLyric Flow Cytometer (BD) using FACS Diva Version (BD, V.8.0.1) software. For high-dimensional analysis, fcs files from individual samples were labelled with the sample ID and disease type and concatenated into one fcs file. In total, 307 430 viable PBMC from 28 samples were used for downstream analysis. Dimensional reduction was performed using t-distributed stochastic neighbour embedding (t-SNE) for 1000 iterations with a perplexity of 30. Intact cells were gated according to the forward and side scatter areas. After Side light scatter-A/Side light scatter-H deflection, FVS510 was used to remove the dead cells. Hierarchical consensus clustering was performed on the T and B cells using the FlowSOM algorithm. Lymphocytes were gated on CD45-SSC, and CD19+B cells were identified below the lymph gate. Transitional B cells and CD38dimCD24+cells were gated in the CD38 and CD24 plots, respectively. Within the CD38dimCD24+gate, naïve B cells, non-class-switched cells and class-switched cell populations were identified using IgD-CD27 gating. The manually gated populations were overlaid on the t-SNE plots for visualisation. FlowJo V.10.5.1 was used for manual biaxial gating of individual fcs files and the concatenated file for quality control and independent examination.
Isolation of total memory B cells and co-culture with autologous CD8+ T cells
CD19+ CD27+ memory B (MB) cells were isolated using the EasySep Human Memory B Cell Isolation Kit (STEMCELL Technologies, Canada) according to the manufacturer’s instructions. CD8+T cells were isolated using a CD8+ T Cell Isolation Kit (Miltenyi, Germany). MB cells (1×105 /mL) from SLE and Healthy donors were cultured alone or in the presence of autologous CD8+ T cells (4×105 /mL) at 1:4 ratio for 3 days at 37°C. Co-cultured cells were incubated in complete RPMI 1640 supplemented with 10% fetal bovine serum (Invitrogen) supplemented with 25 mM HEPES (Invitrogen), 2 mM L-glutamate (Invitrogen) and 1% non-essential amino acids (Life Technologies). During incubation, cells were stimulated with class B CpG oligodeoxynucleotides (InvivoGen) at 5 µg/mL, 1 µg/mL anti-CD3/CD28 (BD Biosciences, USA).
Proliferation and apoptosis assays of activated memory B cells in vitro
Stimulated MB cells were co-cultured with CD8+ T cells for 3 days. Cultured cells were transferred to flow tubes and incubated with EdU (25 µM) for 30 min. After washing with PBS, cells were incubated with Apollo staining reaction solution and APC-anti human CD19 antibody at room temperature for 10 min in the dark. After washing, the cells were suspended in 200 µL PBS and analysed by flow cytometry using an FACSCalibur (BD Biosciences, San Diego, California, USA). The annexin V-positive MB cells were examined for apoptosis.
The MB proliferation inhibition rate was calculated as follows:
The MB apoptosis index was calculated as follows:
The IgG secreted MB cells by ELISpot analysis
Anti-human IgG monocloning antibodies (MT91/145, Mabtech) were diluted with PBS (PH7.4) and coated onto a 96-well polyvinylidene fluoride plate (Millipore) treated with 70% ethanol for 2 min at a concentration of 15 µg/mL (100 µL/well) overnight at 4°C. The plates were washed with PBS for five times (300 µL/well) and blocked using 10% bovine serum (200 µL/well) at room temperature for 1 hour. MB cells were pre-stimulated with a mixture of R848 at 1 µg/mL and rhIL-2 at 10 ng/mL (Mabtech) for 24 hours. Activated MB cells and CD8+ T cells were inoculated into the plates at a ratio of 1:4 and cultured at 37°C and 5% CO2 for 48 hours. After washing the plate five times with PBS (300 µL/well), 100 µL (1 µg/mL) of biotin-anti human IgG mAb (MT78/145, Mabtech) was added to the plates and incubated at RT for 2 hours. After another round of washing, 100 µL streptavidin-Alkaline phosphatase (1:1000, Mabtech) was added and incubated for 1 hour at RT. 100 µL substrate solution (BCIP/NBT, Mabtech) were added and developed until the distinct spots emerge. The plates were analysed using a CTL reader (ImmunoSpot, Cleveland, Ohio, USA).
Exogenous overexpression of PD-L1 on MB cells
Recombinant lentiviral overexpression of PD-L1 was a gift from Wuhan University. The viral supernatant and medium were mixed in a 1:1 ratio, and the total volume was 1 mL. Polybrene was added at a concentration of 8 µg/mL. MB cells (1×105 cells from patients with SLE) were resuspended in this mixture and inoculated into 24-well culture plates for 24 hours. The expression of PD-L1 on the surface of MB cells was detected by flow cytometry. Transfected MB cells (1×104 transfected MB cells) were cultured with CD8+T cells from healthy donors or patients with SLE at a ratio of 1:4 for 72 hours. Class B CpG oligodeoxynucleotides (InvivoGen) at 5 µg/mL, 1 µg/mL anti-CD3/CD28 (BD Biosciences, USA) were added to the culture as stimulants. IFN-γ expression in CD8+ T cells was measured using flow cytometry.
Statistical analysis
Data analysis was performed using GraphPad Prism V.9.2.0 (GraphPad Software, San Diego, California, USA). Continuous variables were described as medians and calculated using the Mann-Whitney test. Multiple comparisons were performed using the Kruskal-Wallis test after Bonferroni correction. In all statistical analyses, a two-tailed p value<0.05 was considered statistically significant.
Results
Indeterminate QFT is more prevalent in SLE and relevant to more active disease and renal involvement
We compared the QFT results of 329 patients with RMD, including SLE (n=104), primary Sjögren’s syndrome (n=21), systemic sclerosis/mixed connective tissue disease (n=18), idiopathic inflammatory myopathy (n=36), Antineutrophil cytoplasmic antibody-associated vasculitis (n=10), rheumatoid arthritis (RA; n=26), psoriatic arthritis (n=6), ankylosing spondylitis (n=42), polymyalgia rheumatica (n=42), adult-onset still disease (n=11) and idiopathic non-infectious uveitis (n=13) derived from a rheumatology-ophthalmology joint clinic, along with those of 20 healthy donors. The indeterminate QFT result was 17.31% in patients with SLE and 3.56% (p=0.0002) in those with RMDs (figure 1A). Interestingly, patients with SLE with indeterminate QFT results had higher SLEDAI scores versus those with definitive results (10.94 vs 4.02, p<0.0001), and higher anti-double-stranded DNA antibody positive rate (72.22% vs 46.51%, p=0.0472) and titre (109.78 vs 54.21 IU/mL, p=0.0255, normal range <10 IU/mL). More patients with active nephritis (55.56% vs 29.07%, p=0.0306, urinary protein/creatinine ratio>0.5 g) were identified among those with intermediate QFT results (table 1).
QFT results and the differences in IFN-γ release were shown in patients with rheumatic musculoskeletal diseases. (A) The distribution of indeterminate QFT results in patients with rheumatic musculoskeletal diseases (n=329) were showed. (B) The comparison of PHA induced IFN-γ production (P minus N) between non-SLE and SLE using QFT. (C) The messenger RNA expression of IFN-γ in CD4+ and CD8+ lymphocytes were detected by PT-PCR using 2ˆ(−ΔCT) calculation method. AAV, Antineutrophil cytoplasmic antibody associated vasculitis; AOSD, adult onset still disease; AS, ankylosing spondylitis; HC, healthy donors; IFN, interferon; IIM, Idiopathic inflammatory myopathy; MCTD, mixed connective tissue disease; PHA, phytohemagglutinin; PMR, polymyalgia rheumatica; PsA, psoriatic arthritis; pSS, primary Sjögren’s syndrome; QFT, quantitative interferon-γ release test; QFT-N, (negative control); QFT-P, (positive control); QFT-T, (test); RA, rheumatoid arthritis; SLE, systemic lupus erythematosus; SSc, systemic sclerosis. Data were compared by Mann-Whitney test and Kruskal-Wallis test after Bonferroni correction. *p≤0.05, **p≤0.01, ***p≤0.001, ****p≤0.0001.
Baseline characteristics of patients with SLE with indeterminate results versus definitive results of QFT
The indeterminate QFT results in patients with SLE were mainly due to insufficient IFN-γ response to PHA
To further determine the cause of the indeterminate QFT results, IFN-γ levels were analysed in the QFT-T, QFT-N and QFT-P tubes (figure 1A). Although the IFN-γ value in the QFT-N tube was slightly higher in patients with SLE, there was no significant difference (SLE vs non-SLE, mean 0.40 vs 0.21, p=0.0844). Patients with SLE exhibited significantly lower IFN-γ production under PHA stimuli before or after QFT-N normalisation (SLE vs non-SLE, QFT-P: mean 7.33 vs 10.60 IU/mL, p<0.0001; QFT-P minus N: mean 6.92 vs 10.39 IU/mL, p<0.0001) (figure 1B).
Among the 18 indeterminate results in SLE by QFT, three were due to the spontaneous release of IFN-γ (QFT-N>8 IU/mL) and 15 were due to low IFN-γ values in the QFT-P. Such a low readout of QFT-P was attributed to neutralising autoantibodies against IFN-γ in 3 of the 15 cases (online supplemental figure 1). Thus, we deduced that the indeterminate QFT results in the remaining 12 patients with SLE were due to intrinsic T-cell dysfunction, which led to ex vivo IIR to PHA. To further rule out the impact of immunosuppressive treatments, the maximum glucocorticoid exposure within 2 weeks and immunosuppressant exposure within the past 6 months before sampling were compared between the two groups, with the only difference being the mean maximum dose of glucocorticoids, which was higher in IIR-SLE (prednisone equivalent 26.67 vs 14.13 mg/day, and p=0.0244; table 2). To further rule out the impact of glucocorticoids, treatment exposures were analysed in other patients with RMD, with or without indeterminate QFT. First, as a previous report suggested that prednisone equivalent ≥60 mg/day contributes to false-negative readout of the T-SPOT TB assay,10 our samples were below this threshold. Second, no significant difference in glucocorticoid exposure was observed between the patients with RMD with or without indeterminate QFT. Furthermore, glucocorticoid exposure in patients with RMD with indeterminate QFT was comparable to that in patients with SLE with a definite QFT (online supplemental table 2).
The treatment of patients with SLE with and without IIR
To exclude glucocorticoids as a bias for indeterminate QFT results in patients with SLE, the correlation between Glucocorticoids and QFT values was analysed in patients with SLE with indeterminate (R2=0.1735, p=0.0855) or definitive QFT (R2=0.018. p=6981) results (online supplemental figure 2). These data suggest that glucocorticoid dosage at sampling is not directly related to indeterminate QFT readouts. In other words, an indeterminate QFT is likely to be an intrinsic rather than a drug-related effect.
Exhausted CD8+T lymphocytes are responsible for IIR to PHA in patients with SLE
A total of 38 individuals were included, including HDs (n=10), IIR-SLE (n=8) and SLE controls (patients with SLE with definitive negative QTF, n=20). First, the basal messenger RNA (mRNA) expression of IFN-γ produced by CD4+ T lymphocytes in patients with IIR-SLE was comparable to that in SLE controls, although both were higher than those in HDs (HD vs IIR-SLE vs SLE controls, mean 0.75 vs 2.89 vs 3.22, online supplemental table 3). However, the basal IFN-γ mRNA expression in CD8+ T lymphocytes was significantly lower in patients with IIR-SLE and HD than in SLE controls HD versus IIR-SLE versus SLE controls, mean 0.98 versus 1.95 versus 8.33, p<0.001, online supplemental table 3). Under PHA stimulation, CD4+T lymphocyte-induced mRNA expression of IFN-γ was comparable in all three groups in a PHA dose-dependent manner. Strikingly, CD8+ T lymphocytes responded inadequately to PHA concentration (figure 1C, online supplemental table 3). Intracellular staining for IFN-γ at the protein level was performed using flow cytometry (online supplemental figure 3, online supplemental table 3) confirmed that insufficient IFN-γ release in patients with SLE was mainly caused by CD8+ T lymphocytes.
We tested whether CD8+T-cell exhaustion is pivotal for IIR to PHA. We chose a set of reported T-cell exhaustion markers (PD-1, lymphocyte activation gene 3 protein (LAG3), cytotoxic T lymphocyte associated protein 4 (CTLA-4), T-cell immunoglobulin and mucin domain-3, B and T lymphocyte attenuator (BTLA), CD160, CD244 and killer cell lectin like receptor G1 (KLRG1))11–13 to capture the possible gene signature of CD8+ T cells among patients with IIR-SLE (n=5), HD (n=10) and SLE controls (n=15). As expected, exhaustion correlation genes (PD-1, LAG3, CTLA-4 and KLRG1) were specifically upregulated in CD8+ T cells, but not in CD4+ T cells, in IIR-SLE (figure 2A and B, online supplemental table 4). These results were replicated by flow cytometry at the protein level, with the same four exhaustion molecules being upregulated (figure 2C, online supplemental table 4).
The exhaustion markers expressions by T cells in patients with SLE and IIR-SLE. (A–B) Heat map and bar chart were showed the messenger RNA expression of exhausted genes markers in patients with HD (n=10), IIR-SLE (n=5) and SLE (n=15) using RT-PCR. (C) The exhausted genes were shown by flow cytometry in each study group (HD (n=10), IIR-SLE (n=5) and SLE (n=9). BTLA, B and T lymphocyte attenuator; CD, cluster of differentiation; CTLA-4, cytotoxic T lymphocyte associated protein 4; HD, Healthy Donor; IIR, insufficient interferon-γ response; KLRG1, killer cell Lectin like receptor G1; LAG3, lymphocyte activation gene 3 protein; PD-1, programmed death receptor 1; RT, room temperature; SLE, systemic lupus erythematosus. The multiple comparisons were analysed using the Kruskal-Wallis test after Bonferroni correction. *Represent the statistical differences in the comparison with IIR-SLE and HD. #Represent the statistical differences in the comparison between IIR-SLE and SLE. */#, p≤0.05, **/##, p≤0.01, ***/###, p≤0.001, ****/####, p≤0.0001.
The circulating memory B cells were increased along with exhausted CD8+T cells in IIR-SLE
Next, we examined the B-cell populations that interacted with CD8+T cells (the immunophenotyping protocol is provided in online supplemental figure 4. The t-SNE analysis showed that, in parallel with PD-1highCD8+T cells, the proportion of total MB cells, class-switched MB cells and plasmablasts in IIR-SLE (n=8) was significantly higher than that in HD (n=10) and SLE controls (n=10) (figure 3A). Furthermore, we confirmed that PD-L1 expression on the surface of sorted total MB cells was significantly reduced in patients with IIR-SLE (IIR-SLE vs SLE control, Mean Fluorescence Intensity 873.3 vs 2099, p=0.0002) (figure 3B).
The percentage of plasmablasts and memory B cells were increased in the peripheral blood mononuclear cell of patients with IIR-SLE. (A) The dimension reduction analyses of lymphocyte subsets were compared between patients with SLE (8 IIR-SLE, 10 SLE) and 10 healthy donors by flow cytometry. (B) The expression of PD-L1 on the surface of sorted MB cells in three groups was shown by MFI and bar plots using flow cytometry. Multiple comparisons were performed using the Kruskal-Wallis test after Bonferroni correction. *Represents statistical differences in the comparison between patients with IIR-SLE and HD. #Represents the statistical difference between IIR-SLE and SLE. */#, p≤0.05, **/##, p≤0.01, ***/###, p≤0.001, ****/####, p≤0.0001. CD, cluster differentiation; HD, Healthy Donor; ISO, Isotype Control; IIR, insufficient interferon-γ response; MFI, Mean Fluorescence Intensity; MB, memory B; PD-1, programmed death receptor 1; PD-L1, programmed death-ligand 1; SLE, systemic lupus erythematosus.
Exhausted PD1highCD8+T cells were unable to suppress memory B cell and its IgG production in IIR-SLE
Based on the hypothesis that PD1highCD8+T cells might not serve as effective suppressors of autoreactive B cells, MB cells were isolated and co-cultured with autologous CD8+ T cells (HD, n=3; SLE, n=3; IIR-SLE, n=3). As expected, CD8+ T cells in IIR-SLE were incapable of antagonising MB-cell proliferation (inhibition rate, HD vs IIR-SLE vs SLE control, mean 59.82% vs 23.35% vs 60.80%, p<0.005) after 3 days of co-culture (figure 4A and B) (online supplemental figures 5 and 6). In addition, the apoptosis of these MB cells (annexin V-positive) was downregulated in the presence of PD1highCD8+T cells among IIR-SLE (the apoptosis index, HD vs IIR-SLE vs SLE, mean 6.58 vs 1.59 vs 5.16, p=0.0103) (online supplemental figure 7). Accordingly, the transformation of MB cells into antibody-secreting plasma cells was facilitated by the presence of PD1highCD8+T cells from patients with IIR-SLE (figure 4C). Taken together, these results suggest that exhausted CD8+T lymphocytes cannot inhibit autologous MB cells, and the resultant enhanced proliferation and survival of these B cells automatically heightens their transformation into IgG-producing plasma cells in IIR-SLE.
Exhausted CD8+T cells cannot suppress memory B-cell proliferation and survival. (A) Purified memory B cells (1×105/well) were activated with CpG either alone or in the presence of autogenous CD8+T cells. Prior to harvest on day 3, EdU was added to the culture and the proliferation was measured by flow cytometry. (B) The inhibition rate of CD8+T cells on MBs were shown. (C) In the presence of CD8+T cells, the plasma cells transformation and the secretion of IgG were measured using ELISpot. Each group contained three independent samples. The multiple comparisons were analysed using the Kruskal-Wallis test after Bonferroni correction. *Represent the statistical differences in the comparison between MB cells with or without CD8+lymphocytes. *, p≤0.05, **, p≤0.01, ***, p≤0.001, ****, p≤0.0001. CD, cluster differentiation; CpG, Classbcpgoligonucleotide; HD, Healthy Donor; IL, interleukin; IIR, insufficient interferon-γ response; MB, memory B; SLE, systemic lupus erythematosus.
Upregulating PD-L1 in lupus PD-L1low memory B cells reversed IIR in PD1highCD8+T cells
In SLE, PD-L1low B cells are associated with more active disease and T-cell dysfunction.14–16 Interestingly, the PD-L1 retrovirus (online supplemental figure 8)-transfected MB cells (PD-L1trans+MB) from our patients with IIR-SLE were capable of restoring insufficient IFN-γ release by autologous CD8+T cells (MB vs PD-L1trans+MB, mean 13.20% vs 30.30%, p=0.0017) (figure 5A). This indicated a previously unknown two-way crosstalk between PD1highCD8+T cells and PD-L1low B cells in SLE.
IFN-γ production by exhausted CD8+T cell were restored by replenishing PD-L1 or PD-1 agonist or B-cell depletion. (A) The memory B cells overexpressing PD-L1 from patients with IIR-SLE were co-cultured with purified CD8+T cells. The IFN-γ secretion by CD8+T cells using flow cytometry. (B) The IFN-γ production by CD4+ and CD8+cells were detected using flow cytometry in patients with SLE after PD-1 agonist and PD-1 inhibitor stimulation in vitro. (C) Before and after rituximab treatment. Data were compared by Mann-Whitney test and Kruskal-Wallis test after Bonferroni correction. *, p<0.05，**, p<0.01, ***, p<0.001. HD, Healthy Donor; IFN, interferon; IIR, insufficient interferon-γ response; MB, memory B; PD-1, programmed death receptor 1; PD-L1, programmed death-ligand 1; SLE, systemic lupus erythematosu.
PD-1 agonist in vitro and B-cell depletion therapy in vivo rewired the exhausted CD8+ T-cell IFN-γ response
Next, we tested whether PD-1 agonists could rescue exhausted PD-1highCD8+ T cells in IIR-SLE and found that PD-1 agonists could selectively upregulate IFN-γ production in IIR-SLE CD8+ T cells (nil vs PD-1 agonists, mean 11.28% vs 17.44%, p=0.0361), while downregulating IFN-γ production in CD4+T cells. In contrast, in SLE control patients, both CD4+ and CD8+ T-cell IFN-γ responses were suppressed on treatment with PD-1 agonist in vitro. PD-1 inhibitor, on the other hand, increased the production of IFN-γ by both CD4+ and CD8+T cells in SLE controls and only IFN-γ by CD4+ T cells in patients with IIR-SLE. The IFN-γ response remained unchanged in exhausted CD8+ T cells from patients with IIR-SLE under the PD-1 inhibitor challenge (figure 5B).
Finally, we investigated whether B-cell depletion therapy, rituximab (an anti-CD20 monoclonal antibody), could partially restore IFN-γ secretion by CD8+T cells by eliminating overactivated PD-L1low MB cells in patients with IIR-SLE. For the 10 patients with SLE (IIR-SLE, n=4; SLE, n=6; all without B-cell-targeted treatments in the previous 6 months) who received rituximab, samples were collected before and approximately 1 month after the first dose of treatment. IFN-γ secretion by CD4+ T cells remained stable in both IIR-SLE (pre vs post, mean 21.18% vs 17.05%, p=0.1246) and SLE controls (pre vs post, mean 25.89% vs 20.42%, p=0.4029). However, an increase in IFN-γ secretion by CD8+ T cells in IIR-SLE was observed (pre vs post, mean 15.56% vs 22.68%, p=0.0364), as opposed to stable CD8+T-cell IFN-γ production in SLE controls (pre vs post, mean 27.54% vs 20.62%, p=0.1207) (figure 5C).
Discussion
Indeterminate readouts of QFT are more prevalent in patients with SLE than in those with other rheumatic diseases, as validated multiethnically.17–19 This is clinically relevant, as patients with indeterminate QFT often have more active disease (higher SLEDAI), higher anti-dsDNA titres and more active lupus nephritis (LN), implicating QFT’s non-canonical connotations for clinical utility. Therefore, we have attempted to provide an immunological explanation for this phenomenon.
Among the contributing factors of intermediate QFT results, IIR to PHA was shown to be the most likely, such as spontaneous IFN-γ release13 20 and the presence of anti-IFN-γ neutralising autoantibody,21 22 which was narrowed to the specific mechanism underlying IIR.
Further analysis revealed that CD8+T cells bearing exhaustion markers such as PD-1, LAG3, CTLA-4 and KLRG1 were responsible for this in vitro IIR phenomenon. In recent years, T-cell exhaustion has been increasingly recognised in patients with chronic infections.23 Under persistent antigen stimulation, CD8+T cells obtain an exhausted phenotype, characterised by upregulated expression of the aforementioned inhibitory receptors and decreased production of inflammatory cytokines such as interleukin-2, tumour necrosis factor-α and IFN-γ.24–26 However, only a few reports have addressed exhausted T cells in SLE; in one report, T-cell exhaustion was linked to long-term lupus remission.27 Our immunophenotyping data associated exhausted PD-1highCD8+T cells in patients with IIR-SLE with increased level of circulating MB cells and plasmablasts, which is a benchmark of active SLE.28–32 Subsequent T/B cell co-culture experiments confirmed this, as they showed that exhausted PD-1highCD8+T cells could no longer serve as effective B-cell suppressors and the consequent MB-cell overactivation facilitated their transformation into (auto) autoantibody-producing plasma cells. Although a few studies have reported SLE B-cell anergy to CpG stimuli,33 34 this was not the case in our patients with IIR-SLE. Rather, B cells from these patients with IIR-SLE were seemingly more active and ready to proliferate, which is consistent with a clinically and serologically active disease. Nevertheless, the mechanism underlying the dampened inhibitory effect of exhausted CD8+T cells remains unclear, whether through perforin/granzyme-dependent cytotoxicity35 or via an IFN-γ-dependent manner, which requires further investigation.
Taking a closer look, the PD-1 is a T cell-specific molecule structurally and functionally related to CD28. Activated T cells are stimulated by co-stimulatory signals, which upregulate cell surface expression of PD-1.36 Renal immune cell composition analysis by single-cell sequencing in LN revealed low levels of PD-1 in infiltrated CD8+ T cells,37 suggesting that PD-1high CD8+ T cells might only serve as a compromised peripheral checkpoint for immune tolerance. In contrast, PD-L1lowCD27+IgD+ B cells were consistently identified both in the circulation and kidneys of patients with SLE/LN,38 implying that these activated MB cells are effector cells directly relevant to tissue injury. Notably, by focusing on PD-1, we were able to pinpoint that the PD-1/PD-L1 signalling pathway is pivotal to the mutual interaction of T/B cells, which was verified by the significant replenishment of PD-L1 in lupus MB cells can significantly upregulate IFN-γ production by exhausted CD8+T cells.
The clinical implications of the data have been revised. PD-1/PD-L1 blockade (immune checkpoint inhibitors) has become a revolutionary approach in cancer management.39 Conversely, PD-1 pathway agonists may also contribute to tamoxifen autoimmunity. Some studies have suggested that activation of PD-1 on autoreactive lymphocytes can stimulate physiological immunosuppressive pathways and restore immune regulation, making it an appealing treatment for autoimmune diseases.11 12 40 More importantly, a recent phase 2 clinical trial of peresolimab, a stimulator of PD-1, showed good efficacy and safety profiles in patients with active RA.41 Our data uniquely demonstrated that PD-1 agonists might be particularly useful in the context of IIR-SLE, as they can selectively restore exhausted CD8+ T-cell function while downregulating CD4+ T-cell activation. Other appealing therapeutic strategies include PD-L1 manipulation targeting lupus MB cells or B-cell depletion therapy (which has been widely used in refractory SLE) to exert an indirect modulating effect on CD8+ T-cell exhaustion. Although previous induction of CD8+T-cell exhaustion has been implicated as a possible therapeutic strategy for autoimmune diseases,24 recent evidence, including ours, indicates that restoring CD8+ T-cell suppressor function may be more promising.42 43
Our study was limited by the relatively small cohort of patients with SLE with insufficient IFN-γ response secondary to CD8+ T-cell exhaustion. Although the QFT might serve as a convenient screening tool to capture T-cell dysfunction in practice, extrapolation of our results should be performed with caution before larger, multiethnic validation studies become available. Additionally, although it has been shown that the impact of background glucocorticoids and immunosuppressant exposure is unlikely to yield intermediate QFT results, prospective controlled experiments are expected. Further investigations are required to fully shape the immunological landscape of this complex disease and to explore the clinical utility of PD-1/PD-L1-mediated crosstalk between CD8+T cells and MB cells in SLE.
Data availability statement
Data are available upon reasonable request. Not applicable.
Ethics statements
Patient consent for publication
Ethics approval
This study involves human participants and was approved by Ethics Committee of Renji Hospital, Shanghai Jiaotong University School of Medicine (Ethics number: IRB# 2017-041). Participants gave informed consent to participate in the study before taking part.
Acknowledgments
We are grateful to Jie Liu from Becton, Dickinson and Company for their assistance with the flow cytometry analysis.
References
Supplementary materials
Supplementary Data
This web only file has been produced by the BMJ Publishing Group from an electronic file supplied by the author(s) and has not been edited for content.
Footnotes
KW and JZ are joint first authors.
Contributors KW and JZ contributed to data collection, data analysis, data interpretation, developed the figures and tables, who should be considered co-first authors. XF, SH, JL, FS, ZX and HY performed the statistical analyses and contributed to data quality control. JY contributed to English language. LC had access to the study data and vouched for the data and analyses. SY accepts full responsibility for the work and/or the conduct of the study, had access to the data, and controlled the decision to publish. All authors contributed to the intellectual content during the drafting and revision of the work and approved the final version to be published.
Funding This research is supported by grants from the Clinical Research Plan of Shanghai Hospital Development Center (Project No. SHDC2020CR1015B) and the Rheumatology Special Research Project of the Jiading District Central Hospital (2021-JZXZDZK-01).
Competing interests None declared.
Provenance and peer review Not commissioned; externally peer reviewed.
Supplemental material This content has been supplied by the author(s). It has not been vetted by BMJ Publishing Group Limited (BMJ) and may not have been peer-reviewed. Any opinions or recommendations discussed are solely those of the author(s) and are not endorsed by BMJ. BMJ disclaims all liability and responsibility arising from any reliance placed on the content. Where the content includes any translated material, BMJ does not warrant the accuracy and reliability of the translations (including but not limited to local regulations, clinical guidelines, terminology, drug names and drug dosages), and is not responsible for any error and/or omissions arising from translation and adaptation or otherwise.